Feature #7546: Change behavior of `Array#slice` for an argument of `Range` class

• Status: Assigned
• Assignee: matz (Yukihiro Matsumoto)
• Target version: -

[alexeymuranov (Alexey Muranov)]

=begin
This is a concrete proposal to "fix" #4541.

It is also related to #7545.
For this proposal to make good sense, i think it would be nice if #7545 was at least partially accepted.

=== Main proposal

I propose (({Array#slice})) with (({Range})) type argument to work as follows:

a = ['0', '1', '2', '3']
a[1..2] # => ['1', '2']
a[-2..-1] # => ['2', '3']
a[2..1] # => ['2', '1']
a[-1..-2] # => ['3', '2']
a[-1..1] # => ['3', '0', '1']
a[1..-1] # => ['1', '0', '3']
a[1..1] # => ['1']
a[1...1] # => []
a[4..4] # => [nil]
a[4...4] # => []
a[9..9] # => [nil]
a[9...9] # => []
a[1..5] # => ['1', '2', '3', nil, nil]

=== Secondary proposal: consider adding new instance methods to (({Array})) to compensate the changed behavior of (({Array#slice}))

If this proposal is accepted, the code "(({a[1..-2]}))" for an array (({a})) will not work as before.
This can be compensated by adding new instance methods to (({Array})).
For example the following ones.

1. (({Array#clip(fixnum, fixnum)})):

['0', '1', '2', '3'].clip(1, 1) # => ['1', '2']

Thus (({a.clip(1, 1)})) would be a replacement for (({a[1..-2]})).

(It looks strange to have to convert a pair of numbers ((m)) and ((n)) into a range (({m..(-1-n)})) to simply ask an array to remove ((m)) elements from the beginning and ((n)) elements from the end.
If #7545 is accepted, then the "(({a[1..-2]}))" syntax for "clipping" an array will make not much sense and maybe will not be possible.)

2. (({Array#from(fixnum)})), (({Array#till(fixnum)})):

a = ['0', '1', '2', '3']
a.from(1) # => ['1', '2', '3']
a.till(1) # => ['0', '1']
a.from(1).till(-2) # => ['1', '2']

In fact, in ((Rails)) (({ActiveSupport})) there are methods (({Array#from})) and (({Array#to})) like this, but unfortunately they do not accept negative indices.

((Remark)). It would also be possible to have (({Array#clip!})), (({Array#from!})), (({Array#till!})).
=end

[drbrain (Eric Hodel)]

• Target version set to 3.0

This will break existing code so I set it to next major.

[marcandre (Marc-Andre Lafortune)]

• Assignee set to matz (Yukihiro Matsumoto)

-5 from me:

• this doesn't solve any real-life problem I can think of

• it will introduce incompatibilities

• those incompatibilities would be very difficult to find by code review/grep/whatever

• this would make array[42..n] not always the same as array[42...n+1]

• what about String#slice?

The goal is not to invent a new language.

[alexeymuranov (Alexey Muranov)]

marcandre (Marc-Andre Lafortune) wrote:

• this doesn't solve any real-life problem I can think of

For me it solves one: the current behavior does not make sense to me, or i do not understand which abstract object is modeled by Range :).

• this would make array[42..n] not always the same as array[42...n+1]

For integer n ≥ 42 it should be the same, otherwise i propose to consider #7545.

• what about String#slice?

First the same, then discard the nil values to get a possibly empty string.

By the way, why would you slice a 5-element array or a 5-letter string by something like 2..42 ? (And how about Array#from ?)

Edited

[phluid61 (Matthew Kerwin)]

alexeymuranov (Alexey Muranov) wrote:

marcandre (Marc-Andre Lafortune) wrote:

• this doesn't solve any real-life problem I can think of

For me it solves one: the current behavior does not make sense to me, or i do not understand which abstract object is modeled by Range :).

At the risk of sounding glib, there is an alternative solution: learn it.

[alexeymuranov (Alexey Muranov)]

In fact, i do not request particularly this part:

a = ['0', '1', '2', '3']
a[4..4]   # => [nil]
a[9..9]   # => [nil]
a[1..5]   # => ['1', '2', '3', nil, nil]

The following alternative, closer to the current behavior, would be fine with me:

a = ['0', '1', '2', '3']
a[4..4]   # => []
a[4...4]  # => []
a[9..9]   # => []
a[9...9]  # => []
a[1..5]   # => ['1', '2', '3']

[naruse (Yui NARUSE)]

• Target version deleted (3.0)

[hsbt (Hiroshi SHIBATA)]

• Status changed from Open to Assigned